Nutrient and Pesticide Concentrations in Water from Chemically

Chapter 17

Nutrient and Pesticide Concentrations in Water from Chemically Treated Turfgrass S. A. Harrison, T. Κ. Watschke, R. O. Mumma, A. R. Jarrett, and G. W. Hamilton, Jr.

Downloaded by CORNELL UNIV on May 17, 2017 | http://pubs.acs.org Publication Date: February 18, 1993 | doi: 10.1021/bk-1993-0522.ch017

Department of Agronomy, The Pennsylvania State University, University Park, PA 16802

Turfgrass was established on sloped (9 to 14%) plots by sodding with Kentucky bluegrass or seeding with one of two commercial mixtures. After establishment, the plots were identically treated with pesticides (pendimethalin, 2,4-D ester, 2,4-DP ester, dicamba, and chlorpyrifos) and fertilizers (Ν, P, and K) in a maintenance program similar to those employed by professional turfgrass managers in the northeastern United States. Irrigation was applied to the plots several days before and after each chemical application at an initial rate of 75 mm hr and 150 mm hr thereafter, for the hydrologic characterization of the slopes and to produce runoff and percolate samples for chemical analyses. Natural precipitation did not produce detectable levels (> 0.6 mm hr ) of runoff during the course of the study, although several events produced nondetectable flows that were sampled for chemical analyses. Irrigation applied at the rate of 75 mm hr was not sufficient to produce runoff from the newly sodded plots. Irrigations of 150 mm hr x 60 min produced average runoff volumes of 0.8%, 13.4%, and 11.6% of that applied to the sodded and two seeded treatments, respectively. No residues of pendimethalin, chlorpyrifos, or the esters of 2,4-D and 2,4-DP were detected in any sample. Mean concentrations of 2,4-D acid, 2,4-DP acid, and dicamba for individual events ranged as high as 312, 210, and 252 ug L , respectively. Nondetections for these same compounds accounted for 63%, 64%, and 51% of the analyses, respectively, and another 30%, 25%, and 47%, respectively, were below 70 ug L . Highest concentrations were observed in those samples that were collected within several days of application. Nutrient concentrations remained rather constant and generally reflected the nutrient concentration of the irrigation water. Results of this study suggest that runoff quantities and mean concentrations of dissolved pesticides and nutrients in turfgrass runoff and percolate are generally low. -1

-1

-1

-1

-1

-1

-1

0097-6156/93/0522-0191$06.00/0 © 1993 American Chemical Society

Racke and Leslie; Pesticides in Urban Environments ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

PESTICIDES IN URBAN ENVIRONMENTS


192

Contamination of surface and groundwater by agrichemicals (pesticides and fertilizers) can occur through the processes of runoff and leachingfromtreated areas. Chemical transport by these mechanisms is a function of the product chemistry, soil type, rainfall or irrigation, other climatic factors, topography and geology, application rate andtiming,and the biological system into which the chemical is introduced. Most water quality studies in agriculture have focused on food and fiber production systems which account for about 80% of U.S. pesticide use (1). However, the use of pesticides and fertilizers for turfgrass management and the continuing growth of this economic sector (2,3) has increased the need to understand the impact of these chemical applications on associated water quality. Unfortunately, the dissimilar conditions that exist under perennial turfgrass culture versus most conventional agricultural production systems suggest that extrapolation of data between the two is inappropriate. Turfgrass management practices (including chemical inputs) can vary widely, ranging from low maintenance utilityrights-of-wayto highly managed golf putting green surfaces. The bulk of acreage, however, is dedicated to "medium-input" stands for uses such as home lawns, parks, athletic fields, and golf course fairways (4). In his review, "The Fate of Nitrogenous Fertilizers Applied to Turfgrass," Petrovic (5) noted that losses of Ν to leaching were highly variable across studies, ranging between 0 and 84% of the amount applied. These differences between studies reflect variation among factors that influence the N-leaching process. Studies of chemical losses in runoff from turf have been severely lacking, largely due to the fact that most grassed areas have not produced sufficient runoff for sampling purposes. Kelling and Petersen (6) determined nutrient losses in runofffromprivate lawns using simulated rainfall. They observed Ν losses totaling 0.3 to 15.2% of the amount applied, with concentrations ranging between nondetectable and 18.5 mg L . Phosphorous losses were from 0.4 to 11.7%, with concentrations between nondetectable and 8.5 mg L ; and Κ losses were 0.6 to 18.4%, with concentrations between 0.5 to 23.0 mg L . Gross et al. (7) studied nutrient and sediment losses from sodded turf plots receiving five scheduled applications of urea fertilizer per year. Natural precipitation produced such low levels of runoff and nutrient or sediment losses that a separate study was conducted to determine leaching losses. Nitrate-nitrogen concentrations in percolate averaged below 3.5 mg L during the course of the study. Researchers at the University of Rhode Island investigated nitrogen (8) and pesticide (9) lossesfromsloped experimental lawns maintained under high and low input irrigation and chemical management. Runofffromthis site was also a rarity. Mean flow-weighted, inorganic-N concentrations in soil water ranged between 0.2 and 5.6 mg L . Total inorganic-N losses in percolate averagedfrom3.1 to 13.1% of the amounts applied. Concentrations of the herbicides 2,4-D and dicamba rangedfromnondetectable to 15 and 38 ug L , respectively. Distribution of the concentrations was skewed toward lower values, with the vast majority of samples below 1 ug L . Total amounts of 2,4-D and dicamba recovered in percolatefromthe highest loading treatment were 0.4% and 1.0%, respectively, of the amounts applied. In a recent monitoring study of four Cape Cod golf courses, Cohen et al. (10) analyzed groundwater for 17 pesticides and related compounds that are commonly applied to golf courses in the region. Seven of the compounds, including two of which there was no record of use, were never detected. With two exceptions, the remainder were detected infrequently, with concentrations near or below 1 ug L . The exceptions were chlordane 1

1

1

1

1

1

1

1


17. HARRISON ET AL.

193 Nutrient and Pesticide Concentrations in Water

[1,2,4,5,6,7,8,8-octachlor-2,33a,4J Ja-hexahydro-4J-methanoindane], of which there was also no record of use and which was suspected to be a contaminant during well construction; and DBCA (2,4-dichlorobenzoic acid), a compound of unknown origin which may have been an impurity of several pesticide products. The historical lack of water quality data addressing runofffromturfgrass receiving chemical inputs was the major impetus for this research. The objective was to determine the concentration of nutrients and pesticides in runoff and percolatefromtreated turf.


METHODS AND MATERIALS This research was conducted at the Landscape Management Research Center at the University Park campus of The Pennsylvania State University. The site, located on a 9 to 14% slope, was formerly utilized for soil erosion research and was left undisturbed and unattended for nearly 40 yr before being renovated to accommodate this project The soil is a Hagerstown series (Typic Hapludalf) originatingfromlimestone residuum and typical of the karst topography found in dieridgeand valley province of central Pennsylvania (11). The surface soil was classified as clay (23% sand, 36% silt, 41% clay). This textural feature is most typical of subsurface horizons in the Hagerstown series, suggesting that significant erosion of the surface horizon(s) had occurred at the site. Bulk density of the top 15 mm of soil was gravimetrically determined to be 1.24 g cm , in the approximate middle of the common range for fine-textured soils as reported by Brady (12). Depth to bedrock was variable and rangedfrom5 to 60 cm. 3

Turfgrass Plot Preparation. Surface preparation for turfgrass establishment consisted of rototilling (10 cm depth), stone removal, leveling by hand raking, and rolling the 6.5 m by 19 m plots. Plots were separated by a plastic edging material (Edg-King, Oak Brook, IL) that extended 10 cm into the soil to eliminate interplot surface and near surface movement of water or applied chemicals. Each plot contained 21 Weathermatic (Garland, TX) popup sprinkler irrigation heads fitted with #520 nozzles. At the bottom of each sloped plot was an epoxy-coated concrete weir that intercepted runoff water. Runofffromindividual plots was directed through galvanized steel chutes into buildings that housed the flow monitoring and subsampling apparatus. Inside each building, waterfromthe chute flowed through a polyethylene splitting chamber (for subsample collection) and into a partitioned, 0.6 m by 1.2 m by 0.3 m deep steel tank. A length of corrugated plastic pipe (20 cm diameter) was suspended below the splitter to act as a baffle to minimize wave formation in the tank. Water accumulating in the receiving side of the tank flowed into the exit chamber via a 90 degree V-notch in the partition wall and was pumped to a storage/disposal tank. A float and counterweight assembly was positioned in the receiving side of the partitioned tank and was banded to a pulley mounted on a potentiometer. The potentiometer produced a voltage signal that was read every 60 seconds by a microprocessor-equipped datalogger (ACUREX Autocalc Data Acquisition System, Mountainview, CA) in an adjacent laboratory. The water level in the tank was maintained at the bottom of the V-notch so that changing water levels in the tank (a function of runoff flow rate) turned the potentiometer and altered the voltage signal. The datalogger was programmed to convert the voltage signal into a flow rate and record it on a computer tape. The data collection system could be activated manually, or automatically by the detection of rainfall at an adjacent weather station.



194

Prior to seedbed preparation, four pan-lysimeter-type subsurface sampling devices were installed 150 mm below the soil surface to capture percolating water. Two of the samplers were located in the upslope portion and two in the downslope portion of the plots. The subsurface samplers consisted of cylindrical polyethylene containers (265 mm diameter by 150 mm depth) that were filled with 16 mm diameter glass marbles. Their water holding capacity was equivalent to a depth of 38 mm. Polyester geotextile fabric separated the glass ballastfromthe overlying soil and prevented sedimentation in the samplers. Polyethylene fittings installed at the top and bottom of the containers allowed venting and emptying of the samplers. Three turfgrass types (experimental treatments) were established in June of 1985: "CLASSIC," a seed mixture of perennial turfgrass species consisting of 25% 'Merit Kentucky bluegrass (Poa pratensis L A 25% 'Julia' Kentucky bluegrass, 20% 'Shadow' chewings fescue (Festuca rubra ssp. commutata Gaud.), and 30% 'Citation' perennial ryegrass (Lolium perenne L.); "CONTRACT," a seed mixture containing 60% annual ryegrass (Lolium multiflorum Lam.), 20% common Kentucky bluegrass, and 20% creeping red fescue (Festuca rubra L.); and "KBG SOD," a three-year-old Pennsylvania Certified Kentucky bluegrass sod grown on a similar silt loam soil in southeastern Pennsylvaniafromthe following seed mixture: 'Adelphi' (25%), 'Baron' (25%), 'Fylking' (25%), and 'Nassau' (25%). Seed was applied at the rate of 19.6 gm nr (CONTRACT) or 14.9 gm m" (CLASSIC) with a drop spreader and mulched with clean straw. All treatments received a complete fertilizer (according to soil test recommendation) at planting at a rate of 48.6,190.3, and 40.6 kg ha* of Ν, P, and K, respectively. Approximately 90 percent of the Ρ component was applied as superphosphate (46% P2O5) and incorporated during the rototilling operation. The remaining Ρ and the Ν and Κ were added as a 15-15-15 (Ν -P2O5-K2O) fertilizer (N source: ammonium nitrate). The fertilizer was applied and raked into the soil surface just prior to seeding or sodding. Soil pH was 7.0 and no lime was applied.


1

2

2

1

Turf Maintenance Practices. The turf management program employed was typical of professionally treated lawns in the northeastern United States. Plots were mowed weekly to a height of 5 cm and clippings were removed. Irrigation was not employed as a routine maintenance practice, however, scheduled irrigations were used to produce runoff and percolate samples. To avoid disruption of the soil surface and subsequent effects on runoff characteristics, mechanical cultivation techniques such as core aeration, slicing, or spiking were not practiced. The area was not subject to foot or wheel traffic (compaction factors) beyond that required for the maintenance of the plots and experimental procedures. Pesticides included in the study were pendimethalin [N-(l-ethylpropyl)-3,4-dimethyl2,6-dinitrobenzenarnine]; 2,4-D butoxyethanolester [(2,4-dichlorophenoxy) acetic acid]; 2,4-DP butoxyethanolester [2-(2,4-dichlorophenoxy) propionic acid]; dicamba [2-methoxy-3,6-dichlorobenzoic acid]; and chlorpyrifos [Ο,Ο-diethyl 0-(3,5,6-trichloro-2pyridyl)-phosphorothioate]. Beginning in 1986, pesticides and fertilizers were applied to the plots with a Chemlawn (Columbus, OH) spray gun delivering 1640 L ha- . Applications were made four times annually as follows: 1


17. HARRISON ET AL.

195 Nutrient and Pesticide Concentrations in Water 1

SPRING - Pendimethalin (1.68 kg ha ) for preemergence control of annual grassy weeds, plus a complete, soluble fertilizer (36.5,4.2, and 8.0 kg ha- of Ν, P, and K, respectively); (fertilizer analysis (N only): 8.8% nitrate-N, 1.0% ammoniacal-N, 10.2% urea-N) EARLY SUMMER - 2,4-D ester, 2,4-DP ester, and dicamba (1.12,1.12, and 0.28 kg ha ) for postemergence control of broadleaf weeds, plus urea fertilizer (36.5 kg Ν ha ); LATE SUMMER - 2,4-D ester, 2,4-DP ester, and dicamba (1.12,1.12, and 0.28 kg ha- ), plus chlorpyrifos (1.12 kg ha- ) for the control of insect pest species, plus urea (36.5 kg Ν ha- ); FALL - 2,4-D ester, 2,4-DP ester, and dicamba (1.12,1.12, and 0.28 kg ha- ), plus urea (36.5 kg Ν ha ). 1

1

1

1

1

1

1


1

Irrigations were generally conducted one week prior to and two days after each chemical application. Irrigation duration was 90 minutes for pre-application events and 60 minutes for post-application events. In addition, all natural precipitation events were monitored for the occurrence of runoff and percolate. Sampling and Analysis. Runoff water for quality analyses was subsampled continuously from the splitting chamber during each runoff event. Water was transferred at a rate of 16 mL min to a Nalgene (Rochester, NY) one liter high density polyethylene bottle through 0.635 mm diameter Tygon tubing via a peristaltic pump (Masterflex Model No. 7017, Cole Palmer Instrument Co., Chicago, EL). Percolate samples were withdrawnfromthe subsurface basins through Tygon (Norton Performance Plastics, Akron OH) tubing, located in a 10 cm access tube adjacent to and downslopefromthe samplers, with a centrifugal pump. No effort was made to characterize the adsorption of pesticides or nutrients onto the various components of the runoff or percolate collection system. Runoff and percolate were collected immediately following irrigation or within 12 hr of natural precipitation events. A 250 mL aliquot was filtered through a Millipore (Bedford, MA) 0.45 um membrane filter and refrigerated at 4 C until nutrient determinations were completed. The remainder of each sample was frozen until pesticide concentrations were determined by the Pesticide Research Laboratory at Penn State University. Nutrient concentrations were determined colorimetrically with a Hach (Loveland, CO) DR/3000 spectrophotometer. Nitrate- and nitrite-N were determined using a cadmium reduction procedure, and soluble phosphorus as orthophosphate via amino acid reduction of molybdophosphoric acid. Both procedures are standard methods (13) adapted for the Hach apparatus, and are described in the instrument reference manual (14). Potassium was determined via the tetraphenylborate method as described in the Hach (14) reference manual. A procedure was developed by the Pesticide Research Laboratory (15) for the analyses of seven pesticides in two high pressure liquid chromotography (HPLC) runs. The compounds were solid-phase (C ) extracted from lOOmL of sample water and adjusted to pH 2.2. Dicamba and the parent acids of 2,4-D and 2,4-DP were analyzed by ion-pair HPLC utilizing a mobile phase of methanol-water and UV detection at 230 nm. 1

8


196



Pendimethalin, chlorpyrifos, and the two esters 2,4-DIOE and 2,4-DPBEE were quantified with HPLC utilizing a Qg column, methanol-water mobile phase, and UV detection at 230 nm. Total vegetation and weed cover were visually estimated to determine whether stand quality was related to runoff response. Vegetative cover was determined as a percent of the total area and reflects the amount of exposed soil. Weeds were also assessed as a percent of the total area. Statistical Design. The three treatments were arranged in a randomized complete block design with three replications. The three physically fixed water sampling locations (upper and lower lysimeters and runoff) did not allow randomization of this main effect, and resulted in a "split-block type" statistical arrangement of these data. Mean separations for hydrologie data were determined using Fisher's lsd (16). RESULTS AND DISCUSSION Hydrology. During the period November 1985 to December 1988, natural precipitation did not result in detectable levels (> 0.6 mm min ) of runoff. This observation concurs with recent experiments in Maryland (7) and Rhode Island (8,9). However, those studies were conducted on sandy loam soils with moderate slopes of 5 to 7% and 2 to 3%, respectively. In contrast, the site used in this study was located on a borderline claytextured soil with slopes of 9 to 14 percent, and the scarcity of runoff from natural events was unexpected. Several natural events did produce runoff sufficient for sampling purposes, however, the flow rates were below the detection limit of the equipment Initial testing of the irrigation system, originally designed to deliver 75 mm hr , on the bare plots (prior to turfgrass establishment in 1985) produced quantities of runoff that exceeded the capacity of the data collection system. However, this irrigation intensity was not sufficient to produce runoff from newly sodded plots, and the system was refitted to double the original intensity. According to a statistical treatment of historical Pennsylvania rainfall data by Aron et al. (17), the simulated events of 1985 (75 mm hr χ 1 or 1.5 hr) had a return frequency of >100 years in State College, PA, a town neighboring the experimental site. The dissimilar droplet patterns and subsequent impact energies of sprinkler irrigation versus natural rainfall could have been considered in determining the statistical significance of these simulated events or in predicting runoff from such storms. However, since no sediment was present in either the natural or irrigated runoff samples, it was assumed that the turfgrass canopy absorbed most of the water droplet impact energy, thus eliminating this concern. Statistical analyses were conducted for individual irrigated events only (Table I). Varying irrigation intensities and durations and environmental conditions precluded analysis across dates. Significant differences in runoff volume between cover types (seed versus sod, generally) most commonly occurred when runoff volumes were highest. A seasonal pattern was noted with highest runoff volumes most prevalent during the months of September, October, and November. Hydrologie trends and comparisons among treatments were evaluated using data from the 1 hr χ 150 mm hr events. The statistical frequency of a natural storm of this magnitude is not well defined by Aron (17) or other available sources, however, a storm 1

1

1

1


17. HARRISON ET AL.

Table I.

DATE

Nutrient and Pesticide Concentrations in Water

Runoff volumes for irrigated events conducted between November 1985 and December 1988 IRRIGATION

RUNOFF VOLUME

OF


EVENT

% of total applied mm/hr

hr

CLASSIC

CONTRACT

KBGSOD

AVG

1985: 13-Nov 25-Nov

75 75

1.5 1.0 avg

19.3 21.0 a** 20.2

2.0 19.4 a 10.7

0.0 0.0 b 0.0

7.1 ns* 13.4 10.3

1986: 25-Aug 08-Sep 08-Oct 23-Oct

150 150 150 150

0 1.0 1.0 1.0 avg

8.0 9.1 a 16.4 a 18.4 a 13.0

6.5 6.7 a 18.2a 15.4 a 11.7

0.0 0.0 b 0.4 b 2.0 b 0.6

4.8 ns 5.3 11.7 11.9 8.4

1987: 03-Jun 25-Jun 10-Jul 22-Sep 01-Oct 04-Nov 16-Nov 10-Dec

150 150 150 150 150 150 150 150

1.0 1.5 1.0 1.0 1.0 1.5 1.0 1.5 avg

1.5 3.6 4.8 11.7 a 12.5 a 3.3 ab 13.8 a 12.9 a 8.0

5.8 5.2 5.7 11.1a 13.4 a 8.9 a 12.4 a 11.9 a 9.3

0.0 0.1 0.0 0.5 b 2.2 b 0.0 b 4.9 b 4.7 b 1.6

2.5 ns 3.0 ns 3.5 ns 7.8 9.4 4.1 10.4 9.8 6.3

1988: 23-May 05-Jul 11-Jul 22-Aug 27-Aug 10-Oct 19-Oct 03-Dec

150 150 150 150 150 150 150 150

1.3 1.5 1.0 1.3 1.0 1.5 1.0 1.5 avg

8.7 a 0.3 0.6 0.7 1.0 1.5 32.4 8.9 a 6.8

7.4 a 1.2 0.0 0.7 1.1 2.3 49.4 2.6 b 8.1

0.0 b 0.1 0.0 0.0 0.0 0.0 0.0 0.2 b 0.04

5.4 0.5 0.2 0.5 0.7 1.3 27.3 3.9 5.0

ns ns ns ns ns ns

# Treatment means are not different at the (. 10) level of significance ## Treatment means in the same row followed by the same letter are not different at the (.10) level of significance (Fisher's lsd)


197


198


delivering the same depth in 24 hours has a return frequency of approximately 100 years in this region (18). The 1.5 hour events were conducted for the sole purpose of ensuring runoff for sampling. They were not incorporated into the hydrologie analyses because their magnitude and return frequency were on the scale of a natural disaster. In addition, the 60-minute, post-application events were usually conducted 7 to 10 days after the preapplication (90 minute) irrigations, providing more uniform antecedent soil moisture conditions under which to make comparisons. In general, lowest runoff volumes were associated with the sodded plots and seasonal effects that we attributed to soil moisture conditions. Nearly equivalent runoff volumes from the seeded treatments in both years, despite a 100% increase in irrigation intensity and volume, suggest that a substantial increase in the infiltration capacity of these plots occurred between the spring of 1985 and fall of 1986. This corresponds to a rapid increase in vegetative cover in the seeded plots during the two growing seasons following establishment of the plots. Turfgrass cover and weed competition were noticeably different between the two seeded treatments until mid-1987. The CONTRACT treatment was quicker to establish than CLASSIC and competed more effectively with annual weeds in 1985 because of the high percentage of annual ryegrass in the mixture. The situation reversed in the following year, however, when the perennial species in the CLASSIC mixture aggressively competed with weeds but the CONTRACT plots suffered a loss of plant density from winter kill of the annual ryegrass component KBG SOD, on the other hand, provided an "instantaneously" stable perennial grass population of very high density. The effect of establishment method on overland flow patterns was such that average runoff volumes from 1 hr irrigations in 1986-7 were 0.8%, 13.4%, and 11.6% for KBG SOD, CLASSIC, and CONTRACT, respectively. Differences in runoff volume reflect a higher infiltration capacity for the sodded plots compared to those that were seeded. This result was probably related to the superior vegetative cover provided by the Kentucky bluegrass sod, especially during the two years following establishment However, differential runoff responses continued after the vegetative stand on the seeded plots stabilized (August 1987), suggesting that turf type or establishment method influenced the runoff process in some manner beyond the effect of areal vegetative ground cover. Two possibilities are stand density and thatch development Although we did not quantify plant density, the rhizomous growth habit of the mature Kentucky bluegrass sod produced a densely matted tangle of plants and thatch that completely covered the soil. The seeded plots contained large proportions of noncreeping, bunch-type grasses and never reached this level of plant density. Even after 5 yr of development bare soil was evident on these plots when the turf foliage was separated by hand. KBG SOD possessed an average 9 mm of compressed thatch at installation and 11.5 mm in 1987, while the seeded plots averaged 4.4 mm in 1987. Thatch was practically nonexistent on much area of the seeded plots, except where sample cores happened to dissect a plant crown. It is likely that these treatments differentially affected soil physical properties which influence the infiltration process. Most notably, sodding completely protected the soilfromthe impact energy of rainfall and irrigation. Conversely, the essentially bare soil of the seeded plots was subject to structural degradation which typically results in the formation of a surface crust that impedes infdtration (19). No experimental measurements were made to confirm this notion.



17. HARRISON ET A L


In each of the 3 yr, runoff volumes tended to be higher in the fall, particularly during the months of October and November. The seasonal variation in runoff volumes may have been associated with increases in antecedent soil moisture levels resulting from seasonal decreases in évapotranspiration rates (20), however, neither évapotranspiration rates or soil moisture conditions were measured. The pattern was especially evident in 1988. Under the severe drought conditions of the summer months, very low levels of runoff were observed from all treatments. In October, when a substantial natural event left the soil in a very wet condition, an irrigated event conducted the following day produced the highest levels of runoff observed during the course of the study. Runoff volumes were substantially lower than anticipated. Introduction of the experimental conditions (land use, soil type, and antecedent soil moisture) into the widely adapted SCS Soil-Cover-Complex method (18) yielded runoff depth estimates of 75 to 90 mm for a 150 mm rainfall (Curve Numbers of 70 to 80), as opposed to the 1.2 to 20.1 mm (Curve Numbers of 30 to 40) observed in the study. The differences between predicted and observed data are even more acute when one considers that the SCS method assumes an S-shaped mass rainfall distribution pattern and a 24-hour storm duration. In this research, a constant intensity, one hour duration irrigation pattern was used, which should have increased the observed runoff depth over the predicted. SCS Curve Numbers reflect site conditions that affect runoff including soil infiltration rate, vegetative cover type, soil conservation practices, and antecedent soil moisture levels. Tables of CN values available in the literature (182022) for common field conditions have been empirically derived from watershed studies of rainfall-runoff relationships and were intended for predicting watershed level impacts of land use/management changes on basin runoff characteristics (23). The results of this experiment support the contention of Rallison and Miller (23) that the use of CN's for less-than-watershed-scale applications is inappropriate, and a more site specific, infiltration based method of determining them is needed for runoff event simulation in field scale applications. Rallison and Miller (23) also noted the inherent weakness of using the CN method in karst regions and other areas of similar subsurface permeability. Ritter (24) described a subcutaneous zone of highly weathered limestone that commonly exists between the soil and the intact bedrock. This zone generally exhibits even greater porosity and permeability than the underlying bedrock, and is very conducive to subsurface storage and transmission of water. Average runoff retention data for these plots suggests that such a zone may have existed at this site. Retention curves for the 60 min events of 1986-7, while significantly higher for KBG SOD than either of the seeded treatments, indicate that steady-state infiltration conditions were approached but often not established for any of the treatments. Water Quality Analyses. A total of 540 samples from the 1986 and 1987 growing seasons were analyzed for pesticides and 648 samplesfromthe 1986,1987, and 1988 growing seasons were analyzed for nutrients. 1

Pesticides. Detection limits for the seven pesticides were as follows: 6.0 ug L for 2,4-D acid; 6.0 ug L for 2,4-DP acid; 20.0 ug L for 2,4-D IOE; 10.0 ug L for 2,4-DP BEE; 2.4 ug L for dicamba; 5.0 ug L* for pendimethalin; and 5.0 ug L- for chlorpyrifos. No residues of pendimethalin, chlorpyrifos, or the esters of the 2,4-D and 2,4-DP were detected in any sample. The first two pesticides are strongly adsorbed to soil particles and 1

1

1

1

1

1


200


organic matter, exhibit marginal water solubility, and are relatively stable in the soil (halflives greater than 30 days) (2526). Under the conditions of this experiment, there was no effective mechanism for aqueous transport. The esters of 2,4-D and 2,4-DP are strongly adsorbed to organic matter (2526) and are readily converted to the parent acids within planttissue(27), so their absence was not unexpected. Mean concentrations of 2,4-D acid, 2,4-DP acid, and dicamba for individual events ranged as high as 312,210, and 252 ug L- , respectively. Such high concentrations were not common, however. Nondetections accounted for 63%, 64%, and 51% of the 2,4-D acid, 2,4-DP acid, and dicamba analyses, respectively. Another 30%, 25%, and 47% of the 2,4-D acid, 2,4-DP acid, and dicamba analyses results, respectively, were below 70 ug L . We arbitrarily note this concentration level because it is the U.S. Environmental Protection Agency's enforceable Maximum Contaminant Level for 2,4-D in public drinking water, and the most limiting of legal standards for the 3 compounds. The conditions under which most of these samples were produced represent an almost unimaginable event All things considered, it is highly unlikely that the number and concentration of these samples is indicative of what might be expected from a similar lawn under natural conditions. Statistical analyses (Fisher's lsd) of the data for these compounds and the three fertilizer nutrients revealed that significant concentration differences between treatments (turf cover types) or sample locations (runoff, percolate) were not common. Of the 29 dates for which nutrient data were available, significant effects were noted on a total of eight dates when all three of the nutrients were considered and on no more than four dates for any individual nutrient For pesticide data, significant effects were noted on a total of six of 21 dates, and on three dates for each individual compound. No clear trends of preferential chemical movement into runoff or percolate, orfromany cover type, were evidentfromthose data where significant effects were noted. Thus, these main effects are not addressed in the remaining discussion. Chemical application dates and treatment-averaged sample concentrations in runoff and percolate are plotted in Tables Π, ΠΙ, IV, and V. Because of the continuous sampling method, runoff concentrations reflect averages for entire events. The 38 mm depth capacity of the subsurface samplers limited observations to this initial flush of percolate. Any additional percolate was assumed to have passed around the sampling units. Generally, highest concentrations of pesticides were observed in samples collected from events occurring within several days of an application. A similar trend was noted by Wauchope (26) in his review of pesticides in agricultural runoff. To date, no other runoff data from highly managed turf are available for comparison to these results. Gold et al. (9) reported their percolate data as flow-weighted, geometric annual mean concentrations rather than individual events, again making comparisons impossible. However, they noted that over 90% of their samples were of concentrations below 5 ug L , regardless of watering practices or pesticide application rate. Although significant concentrations of dicamba and the acids of 2,4-D and 2,4-DP were occasionally observed in our samples, dissipation of these chemicals appears to have been fairly rapid and in line with predicted soil life expectancy (28). 1


1

1

Nutrients. Nutrient concentrations did not exhibit any spécifie patterns in relation to application dates. In fact, they remained rather constant and generally reflected the nutrient content of the water supplying the irrigation system. Average concentrations of


17. HARRISON ET AL.


Table Π. Application dates and rates of active ingredient (italicized) of 2,4-D acid, 2,4-DP acid, and dicamba, and mean concentrations (ug L* ) in runoff 1


EVENT (I = irrigated, Ν = natural)

DICAMBA

2,4-D

2,4-DP

02-Aug-86 09-Sep-86 (I) 09-Oct-86 (I) 21-Oct-86 24-Oct-86 (I) lO-Nov-86 (I) 05-Dec-86 (I)

1.12 kg ha' 46 ND 1.12 kg ha 196 40 ND

1.12 kg ha 24 ND 1.12 kg ha' 130 21 ND

0.28 kg ha 12 ND 0.28 kg ha: 24 19 ND

15-May-87 (I) 03-Jun-87 (I) 15-Jun-87 (N) 25-Jun-87 (I) 06-Jul-87 (N) 07-Jul-87 10-Jul-87 (I) 27-Jul-87 (N) 22-Sep-87 (I) 28-Sep-87 30-Sep-87 (I) 05-Oct-87 (N) 04-Nov-87 (I) lO-Nov-87 17-Nov-87 (I) 10-Dec-87 (I)

ND ND ND ND ND 1.12 kg ha 45 ND ND 1.12 kg ha ND ND ND 1.12 kg ha 12 ND

ND ND ND ND ND 1.12 kg ha* 48 ND ND 1.12 kg ha ND ND ND 1.12 kg ha 40 ND

ND ND ND 4 ND 0.28 kg ha 11 5 ND 0.28 kg ha4 ND 0.28 kg ha 6 ND

1

1

1

1

1

1

1

1

1

ND = Nondetectable


1

1

1

201


202

Table ΙΠ. Application dates and rates of active ingredient (italicized) of 2,4-D acid, 2,4-DP acid, and dicamba, and mean concentrations (ug L* ) in percolate 1



DICAMBA

2,4-DP

2,4-D

1

02-Aug-86 09-Sep-86 (I) 09-Oct-86 (I) 21-Oct-86 24-Oct-86 (I) 10-Nov-86 (I) 05-Dec-86

1.12 kg ha 35 ND 1.12 kg ha 114 11 ND

1.12 kg ha 34 ND 1.12 kg ha 102 12 ND

0.28 kg ha 11 ND 0.28 kg bar 21 4 ND

15-May-87 (I) 03-Jun-87 (I) 15-Jun-87 (N) 25-Jun-87 (I) 06-Jul-87 (N) 07-Jul-87

ND ND ND ND ND 1.12 kg ha 84 ND ND 1.12 kg ha 71 6 ND ND 1.12 kg ha 12 26 ND ND

ND ND ND 22 ND 1.12 kg ha 89 ND ND 1.12 kg ha 105 10 ND ND 1.12 kg ha 41 81 ND ND

ND ND ND 3 ND 0.28 kg ha 22 ND ND 0.28 kg ha* 118 52 26 ND 0.28 kg ha 57 51 41 ND

10-Jul-87 (I) 27-M-87 (N) 22-Sep-87 (I) 28-Sep-87 30-Sep-87 (I) 05-Oct-87 (N) 08-Oct-87 (N) 04-Nov-87 (I) lO-Nov-87 16-Nov-87 (N) 17-Nov-87 (I) 08-Dec-87 (N) 10-Dec-87 (I)

1

1

1

1

1

1

1

1

1

1

1

ND = Nondetectable


1

1

17. HARRISON ET AL.


Table IV. Dates and rates of applied nutrients (Uuikized), and mean concentrations (mg L- ) in runoff 1



Ν 1

02-Aug-86 09-Sep-86 (I) 09-Oct-86 (I) 21-Oct-86 24-Oct-86 (I) lO-Nov-86 (I) 05-Dec-86 (I)

36.5 kg ha 1 2 36.5 kg ha 2 2 2

15-May-87 (I) Ol-Jun-87 03-Jun-87 (I) 15-Jun-87 (N) 25-Jun-87 (I) 06-Jul-87 (N) 07-M-87 10-Jul-87 (I) 27-Jul-87 (I) 22-Sep-87 (I) 28-Sep-87 30-Sep-87 (I) 05-Oct-87 (N)

2 36.5 kg ha 2 ND 1 2 36.5 kg ha 2 4 2 36.5 kg ha 2 ND

24-May-88 (I) 25-May-88 27-May-88 (I) 05-Jul-88 (I) 07-Jul-88 ll-Jul-88(I) 22-Aug-88 (I) 25-Aug-88 27-Aug-88 (I)

2 36.5 kg ha 5 3 36.5 kg ha 3 5 35.5 *g Λα-' 4

1

1

1

1

1

1

Ρ

Κ

-

-

1 ND

3 3

-

-

2 6 1

5 4 1

ND 4.2*gJw-' 1 5 1 1

1 8.0 kg ha 4 3 4 1

-

-

1 2 ND

2 3 1

-

-

1 ND

2 1

1 4.2 kg ha 3 1

1

1

6 8.0 kg ha 19 7

1

-

-

2 3

6 5

-

-

3

8

ND = Nondetectable Racke and Leslie; Pesticides in Urban Environments ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

203

204


Table V. Dates and rates of applied nutrients (italicized), and mean concentrations (mg L' ) in percolate 1



Ν

02-Aug~86 09-Sep-86 (I) 09-Oct-86 (I) 21-Oct-86 24-Oct-86 (I) lO-Nov-86 (I) 05-Dec-86 (I)

36.5 kg ha 3 1 36.5 kg ha 3 ND 1

15-May-87 (I) Ol-Jun-87 03-Jun-87 (I) 15-Jun-87 (N) 25-Jun-87 (I) 06-Jul-87 (N) 07-Jul-87 lO-Jul-87 0) 27-Jul-87 (N) 22-Sep-87 (I) 28-Sep-87 30-Sep-87 (I) 05-Oct-87 (N) 08-Oct-87 (N) lO-Nov-87 08-Dec-87 (N) 10-Dec-87 φ

1 36.5 kg ha 2 ND 1 ND 36.5 kg har 2 1 2 36.5 kg ha 2 2 2 36.5 kg ha ND 2

24-May-88 (I)

1

1

1

_

_

3 2 2 1 1

ND 4.2 kg ha 1 2 1 ND 1 1 ND 1 1 ND ND 1

1 8.0 kg ha3 3 2 2

1 4.2 kg ha 2 1 2 2 ND 2 2 2

1 8.0 kg ha12 3 3 3 2 3 3 4

1

1

1

Κ

1 ND 1 ND ND

1

1

Ρ

1

27-May-88 © 05-Jul-88 (I) 07-Jul-88 ll-Jul-88(I) 19-Jul-88 (N) 22-M-88 (N) 22-Aug-88 φ 25-Awg-M 27-Aug-88 (I) 30-Aug-88 (N)

10 3 36.5 kg ha 4 4 4 4 36.5 kg ha 2 4

1

1

1

-

1 3 1 3 2 2 1 2

ND = Nondetectable


1

17. HARRISON ET AL.


Ν, P, and Κ in the laboratory tap water (same water supply as irrigation) were 2.1,0.4, and 0.4 mg L , respectively, and ranged between nondetectable and 4.1,3.7, and 4.0 mg L , respectively. Water quality standards have not been established for Ρ and K, thus the low concentrations we measured are not discussed further. Unlike nitrate-N, dissolved Ρ and Κ are not generally considered to be a health hazard to humans or wildlife, although elevated levels of phosphate can cause algal blooms and eutrophication of surface water bodies. The absence of elevated concentrations immediately after application is interesting, particularly for Ν which was applied four times annually. Detection of Ν would have hinged on the conversion of urea-N (applied form) to the N0 "- or N0 " -N for which we analyzed. It is likely that a significant amount of the urea was either absorbed by foliage or lost due to NH3 volatilization during the first two or three days following application. Wesely et al. (29) applied urea-N (2.5 g N m ) to eight N-deficient turfgrasses grown under controlled conditions and observed an average of 38.5% absorption of the applied urea by foliage in thefirst48 hours. Volatilization of urea is the result of its hydrolysis to NH in the presence of the enzyme urease. High levels of urease activity in turfgrass thatch and foliage under summer-like conditions have been documented (30). Titko et al. (31) observed volatilization losses from turfgrass grown under simulated field conditions to be as high as 31.4% of dissolved urea applied at the rate of 73 kg Ν ha* . Nitrification of N H to N0 " or N0 " by soil microbes, however, did not likely occur in the initial two days following our applications, since the delivery volume was sufficient to wet the foliage and exposed thatch but not to incorporate the spray solution into the soil. Since neither urea-N or N I V - N were analyzed for in our procedure, the amount remaining is unknown. The fate of these surface-resident Ν forms under the extreme irrigations conducted in the study is also unknown. The apparent absence of NO3-N at later dates could be due to plant uptake of NH4 or N0 ", leaching of dissolved urea below the detection zone (150 mm), or a combination of both. Denitrification may also have played a part However, the irrigation events that would have favored denitrification conditions were isolated and short-lived. 1

1


3

2

-2

3

1

+

4

2

3

+

3

CONCLUSIONS In this study, very low levels of runoff were measured, regardless of the method of turfgrass establishment. Quantities of dissolved nutrients and pesticides in runoff or percolate were also quite small for all treatments. The data reported here are meant to be representative of root zone and "curbside" levels that might be found in a similar suburban yard. They ignore the continued cycling of nutrients and degradation of pesticides, and the dilution effect of other urban and suburban stormwaters that can influence the environmental impact of these compounds on the watershed. These data suggest that under normal rainfall conditions the quantities of dissolved pesticides and fertilizer nutrients in runoff and percolate that are transported from turfed sites are low. It is not clear how well these data reflect the response of turfgrass that is subjected to the cultural and physical stresses common to intensively used or highly trafficked areas, or that is improperly established or poorly managed. According to Aron's (17) treatment of Pennsylvania rainfall data, the 100-year-frequency storm of 60-minute duration has a depth of 58 mm. On this site, a 75 mm depth from an irrigation of the same duration did



206

not produce detectable runoff from sodded slopes. It is the experience of this author that such extreme events can produce surface ponding and runoff from turfed areas. The effects of construction practices, athletic wear, foot and wheel traffic, pest infestations, and the hydrologie condition of adjacent areas all impact the hydrologie character of a given site. Further research into the characterization of home lawns, golf courses, athletic fields, and other turfgrass areas is critical to understanding the implications of these and similar research data.


ACKNOWLEDGMENTS This work was supported in part by the U.S. Department of the Interior, U.S Geological Survey Grant No. 14080001G in cooperation with the Environmental Resource Research Institute, The Pennsylvania State University. The mention of trade names does not constitute their endorsement by the authors or by The Pennsylvania State University. LITERATURE CITED 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

Nielsen, E. G.; Lee, L. K. Agricultural Economic Report Number 576. USDA/ERS. 1987. Anonymous. American Lawn Applicator. 1988, 9(10). Lawn Inst. Harvests; Roberts, E. C., Ed.; The Better Lawn and Turf Institute, Pleasant Hill, TN, 1986. Roberts, E. C.; Roberts, B. C. Lawn and sports benefits; The Better Lawn and Turf Institute. Pleasant Hill, TN, 1988; p. 28. Petrovic, A. M . J. Environ. Qual. 1990, 19(1):, pp. 1-14. Kelling, Κ. Α.; Peterson, A. E. Proc. Soil Sci. Soc. Am. 1975, 39, pp. 348-352. Gross, C. M.; Angle, J. S.; Welterland, M. S.J. Environ. Qual., 1990, 19(3), pp. 663-668. Morton, T. G.; Gold, A. J.; Sullivan, W. M . J. Environ. Qual. 1988, 17(1), pp. 124-130. Gold, A. J.; Morton, T. G.; Sullivan, W. M.; McClory, J. Water, Air, and Soil Pollution. 1988, 37, pp. 121-129. Cohen, S. Z.; Nickerson, S.; Maxey, R.; Dupuy Jr., Α.; Senita, J.A. Ground Water Monitoring Review. 1990, 10(4), pp. 160-173. National Cooperative Soil Survey. Soil survey of Centre County, Pennsylvania. Soil Conservation Service, Wash. D.C., 1981. Brady, N. C. The nature and properties of soils. 8th ed. Macmillan, NY, 1974; pp. 58-62. American Public Health Administration. Standard methods for the examination of water and wastewater. 16th ed. American Public Health Association, NY, 1985. Hach Company. D/R 3000 spectrophotometer instrument manual. 3rd ed. Ha Company, Loveland, CO, 1986. Bogus, E.; Watschke, T. L.; Mumma, R. O. J. Agric. Food Chem. 1990, 38(1), pp. 142-144. Steel, R. G. D.; Torrie, J. H. Principles and procedures of statistics: A biometric approach. 2nd ed. McGraw-Hill, NY, 1980.


17. 17. 18. 19. 20. 21.


22. 23.

24. 25.

26. 27. 28. 29. 30. 31.

HARRISON

Nutrient and Pesticide Concentrations in Water207

ET A L .

Aron, G.; Wall, D. J.; White, E. L.; Dunn, C.N. Water Resour. Bull. 23(3), 1987. U.S. Department of Agriculture. Urban hydrology for small watersheds. 2nd ed. Soil Conserv. Serv. Tech. Rel. 55. Soil Conserv. Serv., Wash. D.C., 1986. Jennings, G. D.; Jarrett, A. R.; Hoover, J.R. Trans. Am. Soc. Agric. Eng. 1988, 31(31), pp. 761-768. Schwab, G. O.; Frevert, R. K.; Edminster, T. W.; Barnes, K.K. Soil and water conservation engineering. 3rd ed. John Wiley and Sons, NY, 1981, pp. 68-71. Leonard, R. Α.; Knisel, W. G.; Still, D. A. Trans.ASAE.,1987, 30(5), pp. 1403-1418. Viessman, W.; Harbough, T. E.; Knapp, J. W. Introduction to hydrology. Educational Publishers, NY, 1972. Rallison, R. Ε.; N. Miller. In Rainfall-runoff relationships; Singh, V. P., Ed., Proc. International Symposium on Rainfall-Runoff Modeling. Water Resources Publications. Littleton, CO, 1982, pp. 353-364. Ritter, Dale F. Process geomorphology. 2nd edition. Wm. C. Brown Publishers, Dubuque, IA, 1986. Goss, D.; Wauchope, R. D. In Pesticides in the next decade: The challenges ahead; Weigmann, D. L., Ed., Proc. of the Third National Research Conference on Pesticides. Virginia Water Resources Research Center. Blacksburg, VA, 1990, pp 471-493. Wauchope, R. D. J. Env. Qual. 1978, 7(4), pp. 459-472. Menzie, C. M. Bur. of Sport Fisheries and Wildlife Special Scientific Report Wildlife No. 127. U.S. Fish and Wildlife Service. Washington, D.C., 1969, p. 109. Weed Science Society of America. Herbicide handbook. 6th ed. Champaign, IL, 1989. Wesely, R. W.; Shearman, R. C.; Kinbacher, E. J. J. Amer. Soc. Hort. Sci. 1985, 110(5), pp. 612-614. Torello, W. Α.; Wehner, D.J. Agron. J. 1983, 75(3), pp. 654-656. Titko, S., III; Street, J. R.; Logan, T. J. Agron. J. 1987, 79(3), pp. 535-540.

RECEIVED December 2,

1992


Nutrient and Pesticide Concentrations in Water from Chemically

Recommend Documents